The position of historians of science with respect to the problem of truth in historical research lies closer to the viewpoint of scientists (physicists, in particular), who consider themselves «extractors of the truth», than to that of some leading philosophers of science in the second half of the twentieth century (Thomas Kuhn, Paul Feyerabend, and others), who abandoned the concept of truth. It is important to distinguish, as did Jose Ortega у Gasset and Marc Bloch, truths of «scientific rationality» (or «physical rationality») from «historical» or «historico-scientific rationality». The author of this essay uses examples from the history of theoretical physics in the nineteenth and twentieth centuries to describe the complex of ideas regarding reality and truth that are typical for historians of science, i. e. «historico-scientific rationality». A constituent part of this complex is the so-called «presentist approach», whereby the present is used as a key to understanding the past. This approach provides a close link between scientific and historico-scientific reasons. The concept of «historical observation» emphasizes empirical facts in the development of science. Historiographical models of the development of scientific knowledge (Kuhn, Imre Lakatos, and others) provide theoretical resources, which historians can test against concrete material. Understanding the process of construction of fundamental scientific theories remains one of the central theoretical tasks of the history of modern science. As an example of the realization of this task, this essay considers the development of the general theory of relativity. It is also important to study secondary lines and dead ends of scientific development, which helps to balance the deficiencies of the presentist approach. For example, unified field theories based on the program of geometrization, although ultimately unsuccessful, proved heuristically important for the development of modern field theory.